Title:	FEATURE-PRESERVING DATA REDUCTION FOR ACCELERATED VOLUME RENDERING OF TIME-VARYING DATASETS
Author(s):	Yi Cao, Guoqing Wu, Huawei Wang
ISBN:	978-972-8939-74-8
Editors:	Yingcai Xiao
Year:	2012
Edition:	Single
Keywords:	Volume Rendering, Visualization, Time-Varying Data, In-formation Visualization
Type:	Short Paper
First Page:	132
Last Page:	136
Language:	English
Cover:
Full Contents:	click to dowload
Paper Abstract:	The time-varying data visualization is a useful way for scientific discovery because it provides scientists a more in-depth understanding of the inherent physical phenomena behind the massive data. But due to irregular data access and the finite memory capacity bottlenecks on the hierarchical storage of the computer system, the interactive rendering ability for large-scale time-varying data are still a major challenge. Data compression strategy can partly solve the problem, but many redundant data will exist in volume data using this method. In this paper, a feature-preserving data reduction scheme based on the information theory is presented and the large-scale time-varying volume rendering can be substantially accelerated. Using the entropy based importance analysis, the featured data, which is preferred by the scientists, can be extracted from the rural data. Then the subsequent data compression and data transfer are directly operated on these extracted data, meanwhile, the remaining nonsense data was discarded during the process. At the same time, a GPU ray-casting volume render is used for fast rendering. Experimental results show that the multi-stage data reduction can greatly speed up the time-varying volume rendering even dealing with the large scale time-varying data.
	Go Back